Yu 1999

Methods	Matched, nested case‐control study Country: China (Taiwan)
Participants	Participants: 4841 men Inclusion criteria: 30 to 65 years of age; HBsAg‐positive or/and HCV‐positive; recruited at two centres: Government Employee Central Clinics or Liver Unit of Chang‐Gung Memorial Hospital Recruitment: August 1988 to June 1992 Outcome assessment: 31 December 1996 Number of cases: Primary liver cancer: 69 (male/female: 69/0) Case definition: incidence Years of follow‐up: 4.5 to 8.3 years Type of selenium marker: plasma
Interventions	d.n.a.
Outcomes	Analysed cases: 69 of 73 cases analysed (reason for non‐inclusion: no sample available) Statistical methods: conditional logistic regression Variables controlled in analysis: age, cigarette smoking, alcohol intake, plasma levels of retinol/alpha‐tocopherol/alpha‐carotene/beta‐carotene/lycopene Variables controlled by matching: age, year and season of sample collection, recruitment clinic
Risk estimates [95% CI]	Reference category: lowest quintile Results: Primary liver cancer highest quintile: OR 0.62 (95% CI 0.21 to 1.86)
Selenium levels in exposure categories	lowest quintile ≤ 124.90 µg/l highest quintile ≥ 162.40 µg/l
Notes

(lower border; upper border) lower and upper border of the 95% CI (if not otherwise specified)

µ                     micro AFP                alpha‐fetoprotein ALT                 alanine aminotransferase ATBC             Alpha‐tocopherol, beta‐carotene cancer prevention study AU                  arbitrary unit BCC               basal cell carcinoma    BMI                 body‐mass‐index BPH               benign prostate hyperplasia CARET          Carotene and Retinol Efficacy Trial CHD               coronary heart disease CI                   confidence interval CIS                carcinoma in situ CVD              cardiovascular disease dl deciliter d.n.a. does not apply DOM             Diagnostic onderzoek mammacarcinoom EVA               Etude du Vieillissement Antériel EPOZ Epidemiologisch onderzoek naar risico‐indicatoren voor hart‐ en vaatziekten FFQ              food‐frequency questionnaire g                    gram HBsAg           Hepatitis B surface antigen HCC hepatocellular carcinoma HCV hepatitis C virus HPFP Hypertension Detection Follow‐up Programme HPFS            Health Professionals Follow‐up Study HR hazard ratio IU international unit l                    litre m                  milli max. adj.       maximally adjusted MHC              Mobile Health Clinic n                   nano NHS               Nurses‘ Health Study NLCS Netherlands Cohort Study NMSC           non‐melanoma skin cancer NPCT            Nutritional Prevention of Cancer Trial n.r. not reported NSAID non‐steroidal antiinflammatory drugs OR Odds ratio p. page p.o.                per os ppm               parts per million PSA prostate‐specific antigen RCT               randomised controlled trial RR relative risk SCC               squamous cell carcinoma SGPT             alanine aminotransferase TIA                 transient ischemic attack UK United Kingdom US             United States of America VITAL             Vitamins and Lifestyle Study ZnTT              zinc turbidity test